Skip to content

Seven entries per benchmark per metric in influx. #227

Closed
@loonatick-src

Description

@loonatick-src

I am trying to understand the output when using the influx format. For a simple benchmark there are seven entries per benchmark and metric combination. E.g. like so (platform columns removed).

metric,units,test,value,test_average,iterations,warmup_iterations,timestamp
Time(userCPU),ns,blackHole,1000,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,1083,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,1166,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,1167,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,1500,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,2042,1613.0,100000,100,2024-01-31T14:17:56Z
Time(userCPU),ns,blackHole,3333,1613.0,100000,100,2024-01-31T14:17:56Z

It is not immediately obvious to me why they are repeated seven times with only the value column changing, but my guess is that this is the seven-figure summary. Is that correct? If yes, perhaps there should be an additional column indicating which percentile it is, or is the user expected to infer that using the relative orders?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions